IP Library › Granted Patent US 11,729,424
Granted Patent B2
US 11,729,424 · App. 17/543,306 · Granted Aug 15, 2023

Visual quality assessment-based affine transformation

Inventors: Kalyan Goswami (Reston, VA); Esmael Hejazi Dinan (McLean, VA); Tae Meon Bae (McLean, VA)
Assignee: Ofinno, LLC
H04N19/61H04N19/136H04N19/159H04N19/176
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 11,729,424
App. No.
17/543,306
Granted
Aug 15, 2023
Kind
B2
Abstract

A decoder may receive, for a block and from a bit stream, an indication of a decoder-side affine transform, a prediction mode, and a residual block. The decoder may generate a compensated prediction of the block. For example, the decoder may generate the compensated prediction of the block based on the residual block and the prediction mode. The decoder may generate, based on the indication and for each of a plurality of affine transform parameters, an affine transformation of the compensated prediction. The decoder may determine an affine transform parameter, from the plurality of affine transform parameters, based on a visual quality of each of the affine transformations of the compensated prediction.

Claims (38)

1. A method comprising:

receiving, for a block and from a bit stream, an indication of a decoder-side affine transform, a prediction mode, and a residual block;

generating a compensated prediction of the block based on:

the residual block; and

the prediction mode;

generating, based on the indication and for each of a plurality of affine transform parameters, an affine transformation of the compensated prediction; and

determining an affine transform parameter, from the plurality of affine transform parameters, based on a visual quality of each of the affine transformations of the compensated prediction, wherein the visual quality of each of the affine transformations is determined without using the block as a reference.

2. The method of claim 1 , wherein the visual quality of each of the affine transformations is determined based on a visual parameter measurement index (VPMI).

3. The method of claim 1 , wherein the visual quality of each of the affine transformations is determined based on a deep Learning for Blind Image Quality Assessment (DeepBIQ).

4. The method of claim 1 , further comprising receiving a displacement vector based on the prediction mode.

5. The method of claim 1 , wherein the prediction mode comprises an inter prediction coding scheme, an intra prediction coding scheme, or an intra block copy prediction coding scheme.

6. The method of claim 1 , wherein the indication of the decoder-side affine transform comprises a flag.

7. The method of claim 1 , wherein the indication of the decoder-side affine transform comprises a difference between two affine transform parameters.

8. The method of claim 1 , wherein the indication of the decoder-side affine transform is comprised in the prediction mode.

9. The method of claim 1 , wherein the indication of the decoder-side affine transform is conditioned on the residual block.

10. A decoder comprising:

one or more processors; and

memory storing instructions that, when executed by the one or more processors, cause the decoder to:

receive, for a block and from a bit stream, an indication of a decoder-side affine transform, a prediction mode, and a residual block;

generate a compensated prediction of the block based on:

the residual block; and

the prediction mode;

generate, based on the indication and for each of a plurality of affine transform parameters, an affine transformation of the compensated prediction; and

determine an affine transform parameter, from the plurality of affine transform parameters, based on a visual quality of each of the affine transformations of the compensated prediction, wherein the visual quality of each of the affine transformations is determined without using the block as a reference.

11. The decoder of claim 10 , wherein the visual quality of each of the affine transformations is determined based on a visual parameter measurement index (VPMI).

12. The decoder of claim 10 , wherein the visual quality of each of the affine transformations is determined based on a deep Learning for Blind Image Quality Assessment (DeepBIQ).

13. The decoder of claim 10 , wherein the instructions, when executed by the one or more processors, further cause the decoder to receive a displacement vector based on the prediction mode.

14. The decoder of claim 10 , wherein the prediction mode comprises an inter prediction coding scheme, an intra prediction coding scheme, or an intra block copy prediction coding scheme.

15. The decoder of claim 10 , wherein the indication of the decoder-side affine transform comprises a flag.

16. The decoder of claim 10 , wherein the indication of the decoder-side affine transform comprises a difference between two affine transform parameters.

17. The decoder of claim 10 , wherein the indication of the decoder-side affine transform is comprised in the prediction mode.

18. A non-transitory computer-readable medium storing instructions that, when executed by one or more processors of a decoder, cause the decoder to:

receive, for a block and from a bit stream, an indication of a decoder-side affine transform, a prediction mode, and a residual block; and

generate a compensated prediction of the block based on:

the residual block; and

the prediction mode;

generate, based on the indication and for each of a plurality of affine transform parameters, an affine transformation of the compensated prediction; and

determine an affine transform parameter, from the plurality of affine transform parameters, based on a visual quality of each of the affine transformations of the compensated prediction, wherein the visual quality of each of the affine transformations is determined without using the block as a reference.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jun 10, 2022
From: GOSWAMI, KALYAN; DINAN, ESMAEL HEJAZI; BAE, TAE MEON
To: OFINNO, LLC
Reel/Frame 060171/0052 →
Continuity (2)
Provisional Application 63121543 · Dec 4, 2020
Related Publication 20220182676A1 · Jun 9, 2022